Website founded by
Milan Velimirović
in 2006

12:23 UTC
ISC 2020



Remember me

Forgot your
Click here!
to create your account if you don't already have one.

Rating lists


MatPlus.Net Forum General overhauling stipulation convention
You can only view this page!
(1) Posted by Kevin Begley [Friday, Jul 9, 2010 17:17]; edited by Kevin Begley [10-07-09]

overhauling stipulation convention

Current help-selfmate (hs#n) = both sides conspire to achieve a position where white has a selfmate in 1.
It makes more sense for black to play first here, since it is first and foremost a help-problem; but, that's another matter.
I aim here for a broader overhaul of convention.

What then would the be the logical definition of a self-helpmate (sh#n, not to be confused with ser-h#n)?
Current self-helpmate (sh#n) = white moves first to force a position where he can helpmate, his own King, in 1 move.

But, this definition exactly matches that of a semi-reflex-mate (semi-r#n)!
Even more interestingly, it reveals an element is missing from the current help-selfmate.
That is, current help-selfmates are actually only semi-help-selfmates.

What would a full-bodied help-selfmate (full-hs#n) look like?

Current full-help-selfmate (full-hs#n) = both sides conspire to achieve a position where white has a selfmate in 1, with the added condition that black must play any selfmate in 1 provided in the course of the solution.
To be consistent with current hs#n, white should move first here (but again, that convention for hs#n could improve).

[aside: this new stipulation is quite interesting, but for now I'll concentrate on a proposed naming convention...]
In conclusion, I suggest that the following convention is more logical:

hs#n should replace our current full-help-selfmates (the new stipulation defined above, except that black should move first).
semi-hs#n should replace our current help-selfmate (except that black should move first).
sh#n should replace our current reflexmates.
semi-sh#n should replace our current semi-reflexmates.
(Read Only)pid=5567
(2) Posted by Siegfried Hornecker [Friday, Jul 9, 2010 17:25]; edited by Siegfried Hornecker [10-07-09]

In a reflexmate white must checkmate if he can so this stipulation can't be replaced so easily. :-)

But maybe I just don't understand your semantics.
(Read Only)pid=5568
(3) Posted by Kevin Begley [Friday, Jul 9, 2010 17:48]; edited by Kevin Begley [10-07-10]

My semantics call simply for renaming reflexmates... for clarity & consistency... other than that, they are entirely unchanged.

New self-helpmate = (replaces) current reflexmate.
New semi-self-helpmate = (replaces) current semi-reflexmate.

New help-selfmates, and new semi-help-selfmates do change.

new help-selfmate = (replaces) current full-help-selfmate (new stip), except that black now moves first. (if white moves first, it's new hs#n.5).
new semi-help-selfmate = (replace) current help-selfmate, except that black now moves first. (if white moves first, it's new semi-hs#n.5).

Err, I think I understand your question better now...
Black is not really compelled in a reflexmate... that's just how they explain this to the newcomers.
[edit: note I am not calling anybody a newcomer here, I'm only suggesting they invented this story about black's compulsion to explain it easily... and everybody repeats it to the point we all miss the real meaning, and its implied direction.]
The better view of the reflexmate mechanism is one step removed.
That is, in a semi-r#n, think of white's aim being to play n-1 moves, such that after any defense, white can play a helpmate in 1 (of self).
That's why it's better to think of this as semi-self-helpmate (self is the play for n-1 moves, goal is helpmate in 1 of self).
The solution will show the helpmate in 1, of course, as if black were compelled to play it, but actually it's the goal that changes.

#n = make n-1 white moves such that, against any defense, white can checkmate (of opponent's King) in 1.
s#n = make n-1 white moves such that, against any defense, white can selfmate (of self) in 1.
h#n = make n black moves, with help from white, such that white has #1.
semi-r#n (new semi-sh#n) = make n-1 white moves such that, against any defense, white has a helpmate in 1 (of self).
r#n (new sh#n) = same as semi-r#n, except with the added condition that black cannot achieve the same goal (black cannot obtain a helpmate in 1).

hs#n, unfortunately, is currently off (on at least three counts):
1) black should start,
2) it should measure out to n black moves, with help from white, such that white has a selfmate in 1,
3) the opponent should not be allowed to achieve the reversed goal (if black can selfmate in 1, the help play leading to it becomes a try).

semi-hs#n should be defined: make n black moves, with help from white, such that white has a selfmate in 1.
hs#n, should be: same as semi-hs#n, but with the added condition that black cannot achieve the same goal (black cannot obtain a selfmate in 1, of black King).
(Read Only)pid=5569
(4) Posted by Kevin Begley [Friday, Jul 9, 2010 20:04]; edited by Kevin Begley [10-07-09]

Lest there be any doubts, let me make clear exactly how our current convention is already a failure, and how it may prevent further developments...

First, consider the hs#2: white1 black2 white2 black2# yes?

Now, what is a hs#1? Nothing! white1 black1# (equivalent of s#1 -- but, this is completely absurd!)
Under my convention, hs#1 = n black moves, in help-play to achieve selfmate in 1: black1 white1 black2# (this makes sense!)

Now, let's consider what happens if somebody makes goals which are yet another step removed...

What is a self-help-selfmate in our current convention?
Well, we can't build on our system, because the goal of help-selfmate in 1 is already a failure.

In my convention, self-help-selfmate is simple:
Self is the play -- white plays n-1 moves to achieve the goal (goal = help-selfmate in 1, which makes perfect sense in my convention).
And, this can be expanded indefinitely.

You can have a help-self-help-self-help-selfmate in 1, and it will make sense (if only for somebody living in the year 3000).
Help is the play for 1 move, and the goal is self-help-self-help-selfmate in 1.
(Read Only)pid=5570
(5) Posted by Juraj Lörinc [Friday, Jul 9, 2010 20:16]

That is nice, but as the current conventions are already strongly rooted into wealth of literature, it is surely too late to change likes of reflex mate or helpselfmate.
(Read Only)pid=5571
(6) Posted by Kevin Begley [Friday, Jul 9, 2010 20:27]; edited by Kevin Begley [10-07-09]


And I still have [old] books which say 1.P-K4 [that notation changed] :)
The rules of chess used to allow promotion to your opponent's pieces [rule books have changed]
Cyclone themes used to go by other names [that convention has changed]

Are we to believe that all hope ends with a stipulation convention, which has lasted only as long as the help-selfmate?
The longer we wait to fix faulty infrastructure, the more work will be required.

Let hs# composers (and their judges) take note: any hs#2 may eventually read as semi-hs#1.5 (only a question of when we right this wrong).
And, I think we all know how one-point-five movers look (hehe, unless that dogma changes too)!
I'd advise hs#2.5's (in the old convention), at a bare minimum... and you may want to switch to full-bodied help-selfmates (the semi- prefex is best avoided, if possible).

A good example of hs#3, using my suggestion...
lower/UPPER case = black/WHITE moves
parenthesis shows the goal (selfmate in 1)

1.a? A 2.b B 3.c (...C 4.d#) but, black has another move at b, which selfmates!
1.e! E 2.f? F 3.g (...G 4.h#) but, black has another move at g, which selfmates!
1.e! E 2.i! I 3.j (...J 4.k#) solves.

With some luck, some of these letters match up nicely, providing good thematic content.
(Read Only)pid=5572
(7) Posted by Kevin Begley [Saturday, Jul 10, 2010 06:07]; edited by Kevin Begley [10-07-10]

For clarity, here is a very trivial scheme for a hs#1.5 (by my new convention), or what some may call full-hs#2 (those paralyzed by -- or in denial about -- the flaws of an outdated convention).

(= 6+6 )

1...a5 2.Sa4 Rh1#
but, 1...axb6 Rh8#!

1...a6 2.Sxa6 Rh1#

I would encourage everybody to petition the PCCC to set a formal deadline, after which point this new convention can be universally adopted (if there are valid objections to my suggestion, let them be voiced here).
This is why I believe -- strongly! -- that PCCC must have direct oversight authority, relying upon experts (problemists, programmers, and variant enthusiasts). [aside: it is also why I view all attempts to undermine PCCC's authority as detrimental to this art form.]

And, I would welcome composers to show better examples of this new stipulation (here in this forum, in Mat Plus publications, and elsewhere).

BTW: a case can be made, I believe either way, as to whether this new stipulation is afforded variations.
That is, the remarkable nature of this "full-helper" is that it *might* allow for what you'd expect to see in a direct-problem.

1...a 2.A aa#
1...b 2.E bb#
1...c 2.C cc#

Even the keys and tries are colorless here (both white or black can begin).

2.A a 3.AA aa#
2.B b 3.BB bb#
2.C c 3.CC cc#

And, this gives rise to all kinds of interesting themes, perhaps even cycles never before seen.
I suspect this could be in violation of help-play spirit, but it is best to remain objective... let the artists make their case.
(Read Only)pid=5573
(8) Posted by Neal Turner [Saturday, Jul 10, 2010 17:13]


Maybe you should tell that to Mr Petkov.
In StrateGems (April/June 2010) he takes the long established (though uncommon) reflex-helpmate (rh#) and renames it as 'helpmate with reflex obligation for Black'.
Then he puts forward his own idea 'help-reflex mate' (hr#) with WBWB moves instead of BWBW as in the rh#, even quoting some reflex-helpmates with colours reversed as help-reflex mates.
Obviously with this re-branding he's putting forward the help-reflex mate as a companion to the help-selfmate where White starts first.
I suppose it's good if it stirs up some interest in this neglected idea, but I couldn't help having some misgivings when I read the article.
(Read Only)pid=5578
(9) Posted by Kevin Begley [Sunday, Jul 11, 2010 07:45]; edited by Kevin Begley [10-07-11]

I had an interesting conversation with Petko concerning this matter, in fact.
I intend to write an article (given the scope, it's looking more like a book) on the subject.

The crux of my theory is to think of stipulations in an iterative, goal-oriented fashion, rather than aim-oriented (final result).
When viewed as goal-oriented, the "aims" are allowed to be sub-stipulations, which are to be solved iteratively down, until no further resolution is possible.

There is no mythology necessary about black being reflexively compelled -- it just spirals down to a final result (p=player, o=opponent):
p-r#n = resist-play (opponent resists whatever the goal), with the goal of p-h#1 (player to move, has a helpmate in 1).
p-h#n = help-play, (opponent conspires to allow whatever the goal), with the goal of o-d#1 (opponent to move has direct mate in 1).
p-d#n = resist-play, with the goal of o-# (opponent to move, their King stands checkmated).

So, it's easy to see that: d#n = directmate, h#n = helpmate, r#n = reflexmate (I'll come to the matter of semi-r# and r#, soon)...

There really is no self-style of play (that is part of the difficulty here -- our terminology is full of overlap & inconsistency).
Instead, define what is a compelled mate in 1 (which proves to be nearly as fundamental as checkmate itself!):

p-c#1 = player to move has no legal option other than p-d#1.
p-s#n = resist-play,with the goal of o-c#1.

Easy to see that: s#n = selfmate.

However, curiously, there are two (not one!) forms of help-play, thus two forms of helpmate:

p-semi-h#n: help-play, counter-goals may be rejected, with the goal of o-d#1.
(This is the conventional helpmate, invented by Max Lange, the counter-goal (b-d#1) can be ignored.)

p-full-h#n: help-play, counter-goals must be played, with the goal of o-d#1.
(Here the counter-goal, b-d#1, will be played, if black is given the opportunity).

Is something this fundamental really a new stip?
Perhaps I'm unaware of the conventional stipulation for this...

[edit: not new, though perhaps undervalued!
Christian Poisson has kindly informed me of a problem which nicely employed this stipulation.

Colin R. Flood
Feenschach, 1964
(= 4+7 )

h#2 reflex black

1...f6 2.d5 Kd2#
but 2.Kxe3#

1.Bf4! Bd4 2.d5 Kc3#
This fact permeates all stipulations built upon help-play (e.g., semi-/full-hs#), or have helpmate as a sub-goal (e.g., semi-/full-r#)!

This is already proving to be a better system -- already I have found one (full-hs#n) new stipulation, maybe two (full-h#n)...
Moreover, unlike the conventional methodology, there are no inconsistencies, no overlaps, no bum stipulations...
All the fundamental stipulations (even some that may not have been known!) can be viewed in a unified, hierarchical progression, each one built with the iterative goal of any formerly derived stipulation.
(Read Only)pid=5579

No more posts

MatPlus.Net Forum General overhauling stipulation convention